Novel Variants of ELP2 and PIAS1 in the Interferon Gamma Signaling Pathway Are Associated with Non-Small Cell Lung Cancer Survival

Author Manuscript Published OnlineFirst on June 3, 2020; DOI: 10.1158/1055-9965.EPI-19-1450 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Novel variants of ELP2 and PIAS1 in the interferon gamma signaling pathway are associated with non-small cell lung cancer survival

Yu Chen Zhao1,2, Dongfang Tang1,2, Sen Yang1,2, Hongliang Liu1,2, Sheng Luo3, Thomas E.

Stinchcombe.1,4, Carolyn Glass1,5, Li Su6, Sipeng Shen6, David C. Christiani6,7, Qingyi Wei1,2,4**

1Duke Cancer Institute, Duke University Medical Center, Durham, NC 27710, USA

2Department of Population Health Sciences, Duke University School of Medicine, Durham,

NC 27710, USA

3Department of Biostatistics and Bioinformatics, Duke University School of Medicine, Durham,

NC 27710, USA

4Department of Medicine, Duke University Medical Center, Durham, NC 27710, USA

5Department of Pathology, Duke University School of Medicine, Durham, NC 27710, USA

6Departments of Environmental Health and Department of Epidemiology, Harvard School of

Public Health, Boston, MA, 02115 USA

7Department of Medicine, Massachusetts General Hospital, Boston, MA 02114, USA

*Correspondence to: Qingyi Wei, Duke Cancer Institute, Duke University Medical Center and

Department of Population Health Sciences, Duke University School of Medicine, 905 South

LaSalle Street, Durham, NC 27710, USA, Tel.: (919) 660-0562, email: [email protected].

Running title: Genetic variants of interferon gamma signaling predicts lung cancer survival

Conflict of interest: The authors declare no potential conflicts of interest.

Downloaded from cebp.aacrjournals.org on October 1, 2021. © 2020 American Association for Cancer Research. Author Manuscript Published OnlineFirst on June 3, 2020; DOI: 10.1158/1055-9965.EPI-19-1450 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Keywords: Non-small cell lung cancer, interferon gamma signaling, PD-L1, STAT3-interacting protein 1, single-nucleotide polymorphism, survival

Abbreviations: IFNγ, interferon gamma; NSCLC, Non-small cell lung cancer; PD-1,

Programmed cell death protein 1; PD-L1, Programmed death-ligand 1; SNPs, Single-nucleotide polymorphisms; GWAS, Genome-Wide Association Study; PLCO, the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial; HLCS, Harvard Lung Cancer Susceptibility Study; OS,

Overall survival; DSS, Disease-specific survival; LD: Linkage disequilibrium; BFDP: Bayesian false discovery probability; eQTL, Expression quantitative trait loci; TCGA, the Cancer Genome

Atlas; ROC, Receiver operating characteristics; ELP2, Elongator acetyltransferase complex subunit 2; STATIP1, STAT3-interacting protein 1; STAT1, Signal transducer and activator of transcription 1; PIAS1, Protein inhibitor of activated STAT1; PML, promyelocytic leukemia protein; EAF, Effect allele frequency; HR, Hazards ratio; CI, Confidence interval; AUC, Area under the receiver operating characteristics curve.

Abstract

Background: Interferon gamma (IFNγ) is a pleiotropic cytokine that plays critical immunomodulatory roles in intercellular communication in innate and adaptive immune responses. Despite recognition of IFNγ signaling effects on host defense against viral infection and its utility in immunotherapy and tumor progression, the roles of genetic variants of the IFNγ signaling pathway genes in cancer patient survival remain unknown.

Methods: We used a discovery genotyping dataset from the Prostate, Lung, Colorectal and

Ovarian Cancer Screening Trial (n=1,185) and a replication genotyping dataset from the

Harvard Lung Cancer Susceptibility Study (n=984) to evaluate associations between 14,553 genetic variants in 150 IFNγ pathway genes and survival of non-small cell lung cancer

(NSCLC).

Results: The combined analysis identified two independent potentially functional single- nucleotide polymorphisms (SNPs), ELP2 rs7242481G>A and PIAS1 rs1049493T>C, to be significantly associated with NSCLC survival, with a combined hazards ratio (HR) of 0.85 [95%

CI= 0.78-0.92, P<0.0001] and 0.87 (0.81-0.93, P<0.0001), respectively. Expression quantitative trait loci analyses showed that the survival-associated ELP2 rs7242481A allele was significantly associated with increased mRNA expression levels of ELP2 in 373 lymphoblastoid cell lines and

369 whole blood samples. The PIAS1 rs1049493C allele was significantly associated with decreased mRNA expression levels of PIAS1 in 383 normal lung tissues and 369 whole blood samples.

Conclusions: Genetic variants of IFNγ signaling genes are potential prognostic markers for

NSCLC survival, likely through modulating the expression of key genes involved in host immune response.

Impact: Once validated, these variants could be useful predictors of NSCLC survival.

Introduction

Lung cancer is one of the most common malignancies both in the US and world-wide, with

228,150 new cases of lung cancer and approximately 142,670 deaths from this disease in the

US in 2019 [1]. Non-small cell lung cancer, mostly adenocarcinoma and squamous cell carcinoma, are the most common histological subtypes, accounting for around 85% of all lung cancer patients [2]. Despite devoted efforts in the treatment over the past decades, lung cancer remains the cause for the highest cancer-related mortality worldwide, with an underwhelming 5- year survival rate of 18.6% between 2008 and 2014 in the US [3]. Conventional treatments for

NSCLC include surgery, chemotherapy, and radiotherapy for its early stages, but the responses to these treatments are heterogeneous [4], likely due to genetic variation among the patients, such as single-nucleotide polymorphisms (SNPs), the most common genetic variants that could affect both short-term response and long-term prognosis of cancer patients; thus, identifying the role of these genetic factors for NSCLC survival may lead to a better understanding of the variability in treatment outcomes [5].

Recently, utilization of the immune system to halt cancer development and tumor progression has been widely recognized [6,7] and immunotherapy is now considered the

“fourth-pillar” alongside the three conventional treatments [8]. Theoretically, immunotherapy either assists the ability of the immune system to target cancer cells directly or stimulate the immune system in a more general matter [9]. In NSCLC treatment, programmed death-ligand 1

(PD-L1) inhibitors, such as pembrolizumab, are often applied to patients with a high PD-L1 expression in tumors, in combination with chemotherapy to improve therapeutic results [10].

Despite recent advances, not all patients respond to immunotherapy [11]. Therefore, it is important to identify survival-related biomarkers for immunotherapy.

To date, genomeǦwide association studies (GWASs) investigating millions of SNPs at the same time have identified few SNPs that are associated with cancer survival, because a hypothesis-free GWAS focuses on the most-significant SNPs or genes with a significant P value

after the stringent multiple testing correction. In reported GWASs on cancer survival, most identified SNPs lack functional annotations, which limits clinical application of these results

[12,13]. As a promising hypothesis-driven method in the post-GWAS era, a biological pathway- based approach has been applied to reanalyze published GWAS datasets and to evaluate the cumulative effect of SNPs across multiple genes in the same biological pathway [14]. Since much fewer SNPs in candidate genes of a particular biological pathway of interest will be included in the analysis, it avoids unnecessary multiple tests for SNPs that may have no apparent biological significance, which improves the overall study power to identify statistically significant and biologically important associations [14].

Interferons (IFNs) are a group of pleiotropic cytokines that play immunomodulatory roles during intercellular communication in innate and adaptive immune responses [15]. IFNs are divided into three types, of which the type II interferon (i.e., IFNγ) in humans, is the most extensively studied [15]. IFNγ is a signaling protein released by Cytotoxic T cells and type I T helper cells in response to inflammatory or immune stimuli [16]. IFNγ primarily modulates the activation of IFN response factor 1 through the JAK-STAT pathway, leading to the activation of a group of secondary responsive genes that up-regulate pathogen recognition, antigen presentation, and inhibit cellular proliferation [17,18]. Despite IFNγ signaling being crucial for activating the immune system, it remains to be determined whether IFNγ has a role in assisting tumor immune evasion, especially in the context of clinical trials [19]. Furthermore, the roles of genetic variants of candidate IFNγ signaling genes and their biological functions in tumor growth or suppression remain unknown.

In the present study, therefore, we hypothesize that genetic variants in genes related to the IFNγ signaling pathway, a critical cascade in the activation of both innate and adaptive immune responses, are associated with the survival of NSCLC patients. We used the existing genotyping data from two previously published GWASs to test this hypothesis.

Materials & Methods

Study populations

We utilized a GWAS dataset of the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer

Screening Trial as the discovery; PLCO is a multi-center dataset of randomized controlled study conducted between 1993 and 2011 [20]. In addition to data on genotyping and survival, the

PLCO dataset of 1,185 NSCLC patients also included blood samples and personal information about smoking status, histologic diagnosis, tumor stage, treatment method and family history collected at enrollment [21]. Blood DNA samples of the participants were genotyped with

Illumina HumanHap240Sv1.0 and HumanHap550v3.0 (dbGaP accession: phs000093.v2.p2 and phs000336.v1.p1) [22, 23]. Each institutional review board of the participating institutions had approved the PLCO trial with a written informed consent from each of the participants permitting use of the collected data.

For replication, we used genotyping data extracted from another GWAS dataset of 984 histologically-confirmed Caucasian NSCLC patients from the Harvard Lung Cancer

Susceptibility (HLCS) Study which began in 1991 [24], where whole blood samples and personal information were collected after diagnosis, and DNA extracted from the blood samples were genotyped using the Illumina Humanhap610̺Quad array. The genotyped data was utilized for imputation with the MACH3 software based on sequencing data from the 1000

Genomes Project [24].

The use of these two GWAS datasets were approved by the Internal Review Board of

Duke University School of Medicine (#Pro00054575) and the dbGaP database administration for the PLCO dataset (#6404 with dbGaP accession: phs000093.v2.p2 and phs000336.v1.p1).

Gene and SNP selection

The genes involved in the IFNγ signaling pathway were selected using the Molecular Signatures

Database (http://software.broadinstitute.org/gsea/msigdb/index.jsp) with the keyword “interferon

AND gamma”. With the removal of 133 duplicated genes, two pseudogenes, one withdrawn gene and one gene located on X chromosome, 150 genes remained as candidates for further analysis (Supplementary Table 2). Imputation with IMPUTE2 and the 1000 Genomes Project data (phase 3) was performed for SNPs within ±500kb flanking regions of these candidate genes. SNPs within the genes and their ±2kb flanking regions were then extracted according to the following criteria: an imputation info score t0.8 (Supplementary Figure 1), a genotyping rate t95%, a minor allelic frequency (MAF) t5%, and a Hardy–Weinberg equilibrium (HWE) t1×10-5, which resulted in 14,553 (1053 genotyped and 13,500 imputed) SNPs for further analyses.

Statistical analyses

The endpoints for analysis included overall survival (OS) and disease-specific survival (DSS). In the single-locus analysis, we used multivariate Cox proportional hazards regression analysis to analyze the association between each of these SNPs and NSCLC survival in an additive model using the PLCO dataset, with adjustment for various clinical variables including age, sex, smoking status, histology, tumor stage, chemotherapy, radiotherapy, surgery and the first four principal components (Supplementary Table 3) by using the GenABEL package of R software

[25]. We then used Bayesian false discovery probability (BFDP) with a cut-off value of 0.80 for multiple testing correction to decrease the probability of potentially false positive results as recommended for SNPs in high linkage disequilibrium (LD) as a result of imputation [26,27]. We assigned a prior probability of 0.10 and detected an upper boundary hazards ratio (HR) of 3.0 for an association with variant genotypes or minor alleles of the SNPs with P<0.05. The chosen

SNPs were then validated afterwards by using the HLCS genotyping dataset. Next, we performed an inverse variance weighted meta-analysis to combine the results of both discovery

and replication datasets, followed by a multivariate stepwise Cox model to identify independent

SNPs through adjustments for clinical variables, the top four principal components, demographic characteristics, as well as 15 previously published SNPs [28]. We used Manhattan

Plots and Association Plots to visualize the locations and LD of the selected SNPs.

We then used the combined genotypes or alleles of the identified SNPs to evaluate their cumulative effects and the Kaplan-Meier (KM) survival curves to show their associations with survival probability, constructed the receiver operating characteristic (ROC) curve and time- dependent area under the curve (AUC) to illustrate prediction accuracy for NSCLC survival [29], and performed the expression quantitative trait loci (eQTL) analyses with the genomic data from the 1000 Genomes Project and the genotype-tissue expression (GTEx) project [30,31]. The bioinformatics functional prediction for the tagging SNPs was then performed with SNPinfo [32]

(https://snpinfo.niehs.nih.gov), RegulomeDB [33] (http://www.regulomedb.org) and HaploReg

[34] (http://archive.broadinstitute.org/mammals/haploreg/haploreg.php). Lastly, the differences in mRNA expression levels were examined in 111 pairs of lung cancer tissues and adjacent normal tissues from the Cancer Genome Atlas (TCGA) database through a paired Student t test. We also assessed the differences in mRNA expression levels in a larger, but not paired, dataset from TCGA (http://ualcan.path.uab.edu), and the KM survival analysis was performed to evaluate the association between the mRNA expression levels and survival probability

(http://kmplot.com/analysis/index.php?p=service&cancer=lung). All statistical analyses were performed with a statistical significance level of P<0.05 by using the SAS software (version 9.4;

SAS Institute, Cary, NC, USA), unless otherwise indicated.

Results

Associations between SNPs in the interferon gamma signaling pathway genes and

NSCLC survival

The overall flowchart of the present study is shown in Figure 1. Basic characteristics of 1,185

NSCLC patients from the PLCO trial and 984 NSCLC patients from the HLCS study have been described in Supplementary Table 1 [28]. After multiple testing correction by BFDP≤0.80, we identified 340 SNPs that were significantly associated with NSCLC OS (P<0.05), of which 48

SNPs in two genes remained significant after further replication by the HLCS genotyping dataset. Subsequently, we performed a combined-analysis of the PLCO and HLCS datasets for these newly identified SNPs. As shown in Table 1, three representative SNPs were determined after considering the LD among the SNPs, one SNP in ELP2 and two SNPs in PIAS1, were found to be associated with a better survival, without heterogeneity observed between the two datasets. Other 45 SNPs in high LD (>0.80) with these three SNPs in the same genes are presented in Supplemental Table 5; these 45 SNPs also feature the same directionality in terms of survival (Supplementary Table 6).

Independent SNPs associated with NSCLC survival in the PLCO dataset

When the three validated SNPs were included in the multivariate stepwise Cox model for the

PLCO dataset only (because the HLCS study dataset did not have individual genotyping data), two SNPs remained independently and significantly associated with survival, even after adjustment for additional 15 previously reported SNPs significantly associated with NSCLC survival from the same PLCO GWAS dataset (Table 2). The results of selected SNPs from

PLCO and HLCS are summarized in two separate Manhattan plots (Supplementary Figure 2a and Supplementary Figure 2b), respectively, and the regional association plot

(http://locuszoom.org/) of each SNP is shown in Supplementary Figure 3 [45].

In the PLCO dataset with complete adjustment for available covariates, patients with the protective ELP2 rs7242481 A (i.e., GA+AA) allele or PIAS1 rs1049493 C (i.e., TC+CC) allele had a better OS and DSS (Ptrend=0.004 and Ptrend=0.009 for ELP2 rs7242481 A, respectively;

Ptrend=0.006 and Ptrend=0.023 PIAS1 rs1049493 C, respectively) (Table 3). In comparison with the GG risk genotype, the ELP2 rs7242481 GA genotype was associated with a decreased risk

of death (HR=0.85, 95% CI=0.73-0.99, P=0.033 for OS and HR=0.84 95% CI=0.72-0.99,

P=0.033 for DSS), and the ELP2 rs7242481 AA genotype was also associated with a decreased risk of death (HR=0.74, 95% CI=0.59-0.94, P=0.012 for OS and HR=0.76 95% CI=0.59-0.97,

P=0.026 for DSS). Similarly, in comparison with the TT risk genotype, the PIAS1 rs1049493 TC genotype was associated with a non-significant better OS and DSS (HR=0.92, 95% CI=0.78-

1.0, P=0.303 and HR=0.91 95% CI=0.76-1.08, P=0.285, respectively), and the PIAS1 rs1049493 CC genotype was associated with a significantly better OS and DSS (HR=0.75, 95%

CI=0.61-0.91, P=0.005 and HR=0.78 95% CI=0.63-0.97, P=0.022, respectively) (Table 3).

Combined effects of the two independent SNPs in the PLCO dataset

We first combined the significant protective genotypes (i.e., ELP2 rs7242481 TA+AA and PIAS1 rs1049493 CC into a genetic score as the number of protective genotypes (NPGs). As shown in

Table 3, the increased genetic score of the NPGs was associated with a better survival in the multivariate analysis in the PLCO dataset (Ptrend<0.0004 for OS and Ptrend=0.002 for DSS). When we dichotomized all the patients into genetic scores of 0-1 and 2 NPGs, the 2-score group had a significantly better survival (HR=0.64, 95% CI=0.52-0.80, P<0.0001 for OS and HR=0.66, 95%

CI=0.53-0.83, P=0.0004 for DSS), in comparison with the 0-1 score group. As shown in Table 3, the increased genetic score of the NPAs was associated with a better survival in the multivariate analysis in the PLCO dataset (Ptrend<0.0001 for OS and Ptrend=0.0006 for DSS). When we dichotomized all the patients into genetic scores of 0-1 and 2-4 NPAs, the 2-4 score group had a significantly better survival (HR=0.77, 95% CI=0.67-0.89, P=0.0004 for OS and HR=0.79, 95%

CI=0.68-0.91, P=0.002 for DSS), in comparison with the 0-1 score group. Because the NPAs was better than NPGs to evenly dichotomize the patients, we used NPAs to facilitate the stratification analysis. We further presented KM survival curves to depict these associations between protective alleles and NSCLC OS and DSS (Figure 2a-2d).

Stratified analysis for associations between NPAs and NSCLC survival

In the stratified analysis by age, sex, smoking status, histology, tumor stage, chemotherapy, radiotherapy and surgery in the PLCO dataset, there were no obvious differences in survival, nor interactions, between the strata of these covariates observed in either NSCLC OS or DSS

(P>0.05 for all strata, Supplementary Table 4).

The ROC curves and time dependent AUC

We assessed the predictive value of the two SNPs with time-dependent AUC and ROC curves at the 60th month (or five-year survival) in the PLCO dataset. Compared with the model of covariates including age, sex, smoking status, histology, tumor stage, chemotherapy, radiotherapy, surgery and the first four principal components, the time-dependent AUC plot with addition of the two independent SNPs did not improve prediction performance of the model at the 60th month. On the other hand, when we performed the time-dependent AUC and ROC curves at the 120th month and combined the two independent SNPs with the 15 previously published SNPs using the same dataset, the prediction performance of the model was improved significantly for both OS and DSS. The AUCs changed from 87.42% to 89.81% (P=0.024) for

OS and from 87.95% to 90.46% (P=0.014) for DSS (Supplementary Figure 7a and 7b).

The eQTL analysis

In the RNA-Seq data of lymphoblastoid cell lines from 373 European descendants available from the 1000 Genomes Project, the ELP2 rs7242481 A allele showed a significant correlation with increased mRNA expression levels of ELP2 in all additive, dominant and recessive models

(P=9.8x10-5, P=0.005 and P<2x10-4, respectively; Figure 3a, Supplementary Figure 8a and

8b); however, there was no significant correlation between the PIAS1 rs1049493 C allele and mRNA expression levels of PIAS1 in all three genetic models (Figure 3d, Supplementary

Figure 8c and 8d). Then, we performed eQTL by using the data of 369 whole blood samples

and 383 normal lung tissue from the GTEx project and found that the rs7242481 A allele remained significantly correlated with a higher expression level of ELP2 in whole blood samples but not in normal lung tissues (P=2.63x10-15 and P=0.141, respectively) (Figure 3c and 3b). The rs1049493 C allele was correlated with a lower expression level of PIAS1 in both normal lung tissues and whole blood (P=0.008 and P=0.0002) (Figure 3e and 3f). Finally, we performed functional prediction for these two independent SNPs utilizing three online bioinformatics tools:

SNPinfo, RegulomeDB, and HaploReg to predict their biological functions as summarized in

Supplementary Figure 4 and Supplementary Table 7.

Differential mRNA expression analysis in target tissues

As shown in Supplementary Figure 5a-5c, in comparison with adjacent normal tissues, tumor tissues had a higher mRNA expression level of ELP2 in lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), and LUAD+LUSC samples combined (P<0.001, P<0.001 and P<0.001, respectively). In the UALCAN (http://ualcan.path.uab.edu) database, the mRNA expression levels of ELP2 were also significantly higher in tumor tissues in LUAD (P<0.001) and in LUSC (P<0.001) (Supplementary Figure 5d and 5e). A higher expression level of ELP2 was associated with improved NSCLC survival as shown by a KM survival curve of lung cancer constructed online (http://kmplot.com/analysis/index.php?p=service&cancer=lung). On the other hand, as shown in Supplementary Figure 6, compared with adjacent normal tissues, tumor tissues had a lower mRNA expression level of PIAS1 in LUAD (P<0.0001), LUSC (P<0.001) and

LUAD+LUSC combined (P<0.001). In the UALCAN (http://ualcan.path.uab.edu) database, the results were also similar, in which the mRNA expression levels of PIAS1 in tumor tissues were lower in both LUAD tissues (P<0.001) and LUSC tissues (P<0.001) in comparison with normal tissues. However, a higher expression level of PIAS1 mRNA was associated with a better

NSCLC survival in the TCGA database.

Discussion

In the present study, we identified and validated two potentially functional and independent

SNPs (i.e., ELP2 rs7242481 and PIAS1 rs1049493) that were significantly associated with the survival of NSCLC in Caucasian populations. We also found that ELP2 rs7242481 A allele was significantly associated with an increased mRNA expression of ELP2 in 373 lymphoblastoid cell lines from the 1000 Genomes Project and an increased mRNA expression level in 369 whole blood samples from the GTEx project. Additionally, we also found that PIAS1 rs1049493 C allele was significantly associated with a lower mRNA expression level of PIAS1 in 383 normal lung tissues and 369 whole blood samples from the GTEx project. These results were consistent with those of the gene expression analysis between paired tumor and adjacent normal tissue samples and survival analysis in the TCGA database. It is worth noting that a significantly higher mRNA expression level of ELP2 was found in tumor tissues in the TCGA database, yet a higher expression level of ELP2 was associated with a better survival using the same data.

ELP2, or STAT3-interacting protein 1 (STATIP1), is located on chromosome 18 and is responsible for encoding the protein ELP2 (elongator protein 2) that makes up the core subunit of the elongator complex, a histone acetyltransferase complex that is highly associated with

RNA polymerase II and various cellular activities [36,37]. In Homo sapiens, ELP2 is mostly known as STATIP1 or StIP1, a well-known STAT3 interactor that is involved in regulating cytokine signal transduction, and an overexpression of STATIP1 inhibits STAT3 activation and nuclear translocation [38]. Studies in the past have also identified STAT3 as a converging point of various oncogenic pathways, and activated STAT3 may trigger tumor progression by directly regulating oncogenic gene expression as well as promote cancer growth via immunosuppression through the inhibition of immune mediators such as pro-inflammatory cytokines [39,40].

In the present study, we found that the rs7242481 A allele was significantly associated with an increased mRNA expression level of ELP2 in lymphoblastoid cell lines and whole blood

tissues; this finding is consistent with the results from other studies, suggesting that the novel genetic variant rs7242481 A allele may affect survival of NSCLC patients through increasing

ELP2 mRNA expression, likely inhibiting oncogenic effects of the activated STAT3 [39,40].

Additionally, we found that mRNA expression levels of ELP2 in tumor tissues from the TCGA database were higher than in adjacent normal tissues in both paired and non-paired samples and were associated with a better survival of NSCLC. These suggest that overexpression of

ELP2 may have occurred as part of our innate immune response to alleviate STAT3 inhibitory effect on immune-stimulating cytokines, resulting in a more rapid activation of the innate immune system [38,39]. According to the ENCODE database, ELP2 rs7242481G>A is located in a DNase I hypersensitive site with highly observable levels of histone modifications in

H3K4Me3 and H3K27Ac acetylation, suggesting that the ELP2 rs7242481 G>A SNP may lead to significantly enhanced transcriptional activities of ELP2.

PIAS1, or protein inhibitor of the activated STAT1, is located on chromosome 15 and is responsible for encoding the E3 SUMO-protein ligase enzyme, which plays a central role as a transcriptional coregulator of STAT1 [41]. Recent studies suggest that the deregulation in SUMO

(small ubiquitin-like modifier)-related pathways may lead to oncogenic transformation of tumor suppressors such as PML (promyelocytic leukemia protein) [41,42]. An interesting finding is that most of the published molecular and functional studies have proposed or identified PIAS1 as an oncogene in NSCLC and other malignant tumors, demonstrating that PIAS1 plays an essential role in degrading PML and that the expression of PML and PIAS1 are inversely correlated in

NSCLC cell lines [42]. However, results obtained from the TCGA database suggest that PIAS1 may be a potential suppressor gene in NSCLC, because its expression was higher in adjacent normal tissues than in tumor tissues in both paired and non-paired samples with additional evidence that a higher expression levels of PIAS1 to be associated with a better survival.

Therefore, additional mechanistic studies are needed to unravel these discrepancies in the association between PIAS1 expression levels and NSCLC survival. A potential explanation to

the difference in results between TCGA analysis and evidence from studies that support the oncogenic role of PIAS1 [42] might be because that the tumor tissues outlined in TCGA also contain fibroblast cells and immune cells, which may alter the resulting observation.

In the present study, we showed that the PIAS1 rs1049493 C allele was associated with lower mRNA expression levels of PIAS1 in whole blood and normal lung tissues, suggesting that the novel rs1049493 C variant allele may affect survival of NSCLC by lowering PIAS1 mRNA expression levels; however, it must be noted that the directionality of PIAS1 rs1049493 C allele yielded mixed results between mRNA levels of PIAS1 and NSCLC survival. Mixed observations on PIAS1 and cancer survival were also observed in other literatures as well as in the KM Plotter for combined histology of lung cancer (Supplementary Figure 6f) [42,46]. As a result, the directionality and mechanisms of PIAS1 in carcinogenesis and patient survival warrant additional molecular validation. According to the ENCODE database, PIAS1 rs1049493

T>C is located in a DNase I hypersensitive site with observable levels of histone modifications in H3K4Me1 acetylation, suggesting that the PIAS1 rs1049493 T>C SNP may lead to altered transcriptional activities of PIAS1.

Currently, there are conflicting outcomes in the IFNγ signaling: to increase an innate immune cell recruitment and type I T helper cell activation but to induce an increased tolerant molecule (i.e., PD-L1) expression in tumor cells to promote immune evasion [11,43, 44]. Also, further functional investigation into the mechanisms of the ELP2 rs7242481 and PIAS1 rs1049493 SNPs are warranted. Furthermore, since both the discovery and replication datasets were from Caucasian populations, our findings may not be generalizable to other ethnic populations. It is also worth noting that the discovery dataset and replication dataset had some differences in distribution of baseline characteristics, leading to fewer significant SNPs being validated. Although the PLCO discovery dataset had a relatively large number of patients, the number of patients in some subgroups were still relatively small, which could cause a reduced statistical power in detecting potential weak effects of other SNPs. Furthermore, genetic

mutation detection and tumor mutation burden [47] have been considered important for developing targeted therapies as the first-line treatments for cancer (i.e., PD-1 or PD-L1 inhibitors) in precision medicine as well as in the prediction of primary resistance to immunotherapy, but such treatment information was not available for further analysis in the present study. For the translational significance, any novel biomarkers for prognosis, such as the SNPs identified in the present study, would be clinically informative, if their effects on tumor phenotypes of NSCLC, such as tumor microenvironment [48], could be evaluated. This suggests that future studies need to collect such important clinical information in the study design, which could establish the associations between these IFNγ-related SNPs and immunotherapy-related NSCLC survival.

The observed directionality of mRNA levels of PIAS1 and NSCLC survival appears to be inconsistent in the biological plausibility, likely due to the diversity of tissues and different patients used in the analysis, as observed in previously published studies [42,46]; therefore, the expression data should be obtained from appropriate and relevant tissue types of the same study populations to avoid discrepant results. Our findings may not have immediate clinical utility, because the magnitude of the HR for clinical significance needs to be much higher and well-verified compared with those from GWAS studies of survival. Once further validated, however, these findings may allow researchers to explore potentially functional SNPs in more rigorous experimental investigation for their biological plausibility and clinical utility.

Funding

Qingyi Wei was supported by the V Foundation for Cancer Research (D2017Ǧ19) and also partly supported by the Duke Cancer Institute as part of the P30 Cancer Center Support Grant (Grant

ID: NIH/NCI CA014236). The Harvard Lung Cancer Susceptibility Study was supported by NIH grants U01CA209414, CA092824, CA074386 and CA090578 to David C. Christiani.

Acknowledgements

The authors thank all the participants of the PLCO Cancer Screening Trial. The authors also thank the National Cancer Institute for providing the access to the data collected by the PLCO trial. The statements contained herein are solely those of the authors and do not represent or imply concurrence or endorsement by the National Cancer Institute. The authors would also like to acknowledge the dbGaP repository for providing cancer genotyping datasets. The accession numbers for the datasets for lung cancer are phs000336.v1.p1 and phs000093.v2.p2.

We wish to thank all the investigators and funding agencies that enabled the deposition of data in dbGaP and PLCO that we used in the present study: The datasets used for the analyses described in the present study were obtained from dbGaP at http://www.ncbi.nlm.nih.gov/gap through dbGaP accession number phs000336.v1.p1 and phs000093.v2.p2. Principal Investigators: Maria Teresa Landi. Genetic Epidemiology Branch,

Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of

Health, Bethesda, MD, USA. Neil E. Caporaso. Genetic Epidemiology Branch, Division of

Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health,

Bethesda, MD, USA.

Funding support for the GWAS of Lung Cancer and Smoking was provided through the

NIH Genes, Environment and Health Initiative [GEI] (Z01 CP 010200). The human subjects participating in the GWAS derive from The Environment and Genetics in Lung Cancer Etiology

(EAGLE) case-control study and the Prostate, Lung Colon and Ovary Screening Trial and these

studies are supported by intramural resources of the National Cancer Institute. Assistance with phenotype harmonization and genotype cleaning, as well as with general study coordination, was provided by the Gene Environment Association Studies, GENEVA Coordinating Center

(U01HG004446). Assistance with data cleaning was provided by the National Center for

Biotechnology Information. Funding support for genotyping, which was performed at the Johns

Hopkins University Center for Inherited Disease Research, was provided by the NIH GEI

(U01HG004438).

PLCO was also supported by the Intramural Research Program of the Division of Cancer

Epidemiology and Genetics and by contracts from the Division of Cancer Prevention, National

Cancer Institute, NIH, DHHS. The authors thank PLCO screening center investigators and staff, and the staff of Information Management Services Inc. and Westat Inc. Most importantly, we acknowledge trial participants for their contributions that made this study possible.

References

[1] Siegel RL, Miller KD, Jemal A. Cancer statistics, 2019. CA Cancer J Clin 2019;69:7-34.

[2] Molina JR, Yang P, Cassivi SD, Schild SE, Adjei AA. Non-small cell lung cancer:

epidemiology, risk factors, treatment, and survivorship. Mayo Clin Proc 2008;83:584-94.

[3] Siegel RL, Miller KD. Jemal A. Cancer Statistics, 2017. CA Cancer J Clin 2017;67:7-30.

[4] Lemjabbar-Alaoui H, Hassan OU, Yang YW, Buchanan P. Lung cancer: Biology and

treatment options. Biochim Biophys Acta 2015;1856:189-210.

[5] Zienolddiny S, Skaug V. Single nucleotide polymorphisms as susceptibility, prognostic, and

therapeutic markers of nonsmall cell lung cancer. Lung Cancer 2012;3:1-14.

[6] Koury j, Lucero M, Cato C, Chang L, Geiger J, Henry D, et al. Immunotherapies: Exploiting

the Immune System for Cancer Treatment. J Immunol Res 2018;2018:9585614.

[7] Pandya PH, Murray ME, Pollok KE, Renbarger JL. The Immune System in Cancer

Pathogenesis: Potential Therapeutic Approaches. J Immunol Res 2016;2016:4273943.

[8] Mizukoshi E, Kaneko S, Telomerase-Targeted Cancer Immunotherapy. Int J Mol Sci 2019;20.

[9] Wang RF, Wang HY. Immune targets and neoantigens for cancer immunotherapy and

precision medicine. Cell Res 2017;27:11-37.

[10] Gong J, Chehrazi-Raffle A, Reddi S, Salgia R. Development of PD-1 and PD-L1 inhibitors

as a form of cancer immunotherapy: a comprehensive review of registration trials and

future considerations. J Immunother Cancer 2018;6:8.

[11] Borcoman E, Kanjanapan Y, Champiat S, Kato S, Servois V, Kurzrock R, et al. Novel

patterns of response under immunotherapy. Ann Oncol 2019;30:385-96.

[12] Fridley BL, Biernacka JM. Gene set analysis of SNP data: benefits, challenges, and future

directions. Eur J Hum Genet 2011;19:837-43.

[13] Witte JS. Genome-wide association studies and beyond. Annu Rev Public Health

2010;31:9-20.

[14] Gallagher MD, Chen-Plotkin AS. The Post- GWAS Era: From Association to Function. Am J

Hum Genet 2018;102:717-30.

[15] Zaidi MR, Merlino G. The two faces of interferon-gamma in cancer. Clin Cancer Res

2011;17:6118-124.

[16] Lloyd CM, Hessel EM. Functions of T cells in asthma: more than just T(H)2 cells. Nat Rev

Immunol 2010;10:838-48.

[17] Ivashkiv LB, Donlin LT. Regulation of type I interferon responses. Nat Rev Immunol

2014;14:36-49.

[18] Tau G, Rothman P. Biologic functions of the IFN-gamma receptors. Allergy 1999;54:1233-

51.

[19] Castro F, Cardoso AP, Goncalves RM, Serre K, Oliveira MJ. Interferon-Gamma at the

Crossroads of Tumor Immune Surveillance or Evasion. Front Immunol 2018;9:847.

[20] Hocking WG, Hu P, Oken MM, Winslow SD, Kvale PA, Prorok PC, et al. Lung cancer

screening in the randomized Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer

Screening Trial. J Natl Cancer Inst 2010;102:722-31.

[21] Oken MM, Marcus PM, Hu P, Beck TM, Hocking W, Kvale PA, et al. Baseline chest

radiograph for lung cancer detection in the randomized Prostate, Lung, Colorectal and

Ovarian Cancer Screening Trial. J Natl Cancer Inst 2005;97:1832-39.

[22] Mailman MD, Feolo M, Jin Y, Kimura M, Tryka K, Bagoutdinov R, et al. The NCBI dbGaP

database of genotypes and phenotypes. Nat Genet 2007;39: 1181-86.

[23] Tryka KA, Hao L, Sturcke A, Jin Y, Wang ZY, Ziyabari L, et al. NCBI's Database of

Genotypes and Phenotypes: dbGaP. Nucleic Acids Res 2014;42:975-79.

[24] Zhai R, Yu X, Wei Y, Su L, Christiani DC. Smoking and smoking cessation in relation to the

development of co-existing non-small cell lung cancer with chronic obstructive pulmonary

disease. Int J Cancer 2014;134:961-70.

[25] Aulchenko YS, Ripke S, Isaacs A, van Duijn CM. GenABEL: an R library for genome-wide

association analysis. Bioinformatics 2007;23:1294-96.

[26] Park JH, Geum DI, Eisenhut M, van der Vliet HJ, Shin JI. Bayesian statistical methods in

genetic association studies: Empirical examination of statistically non-significant Genome

Wide Association Study (GWAS) meta-analyses in cancers: A systematic review. Gene

2019;685:170-78.

[27] Wakefield J. A Bayesian measure of the probability of false discovery in genetic

epidemiology studies. Am J Hum Genet 2007;81:208-27.

[28] Tang, D, Zhao YC, Qian D, Liu H, Luo S, Patz EF, et al. Novel genetic variants in HDAC2

and PPARGC1A of the CREB-binding protein pathway predict survival of nonsmall cell lung

cancer. Mol Carcinog 2020;59:104-15.

[29] Chambless LE, Diao G. Estimation of time-dependent area under the ROC curve for long-

term risk prediction. Stat Med 2006;25:3474-86.

[30] Consortium GT, Human genomics. The Genotype-Tissue Expression (GTEx) pilot analysis:

multitissue gene regulation in humans. Science 2015;348:648-60.

[31] Lappalainen T, Sammeth M, Friedlander MR, Hoen PA, Monlong J, Rivas MA, et al.

Transcriptome and genome sequencing uncovers functional variation in humans. Nature

2013;501:506-11.

[32] Xu Z, Taylor JA. SNPinfo: integrating GWAS and candidate gene information into functional

SNP selection for genetic association studies. Nucleic Acids Res 2009;37:600-05.

[33] Boyle AP, Hong EL, Hariharan M, Cheng Y, Schaub MA, Kasowski M, et al. Annotation of

functional variation in personal genomes using RegulomeDB. Genome Res 2012;22:1790-

97.

[34] Ward LD, Kellis M. HaploReg v4: systematic mining of putative causal variants, cell types,

regulators and target genes for human complex traits and disease. Nucleic Acids Res

2016;44:877-81.

[35] Wang Y, Liu H, Ready NE, Su L, Wei Y, Christiani DC, et al. Genetic variants in ABCG1 are

associated with survival of nonsmall-cell lung cancer patients. Int J Cancer 2016;138:2592-

2601.

[36] Dong C, Lin Z, Diao W, Li D, Chu X, Wang Z, et al. The Elp2 subunit is essential for

elongator complex assembly and functional regulation. Structure 2015;23:1078-86.

[37] Winkler GS, Petrakis TG, Ethelberg S, Tokunaga M, Erdjument-Bromage H, Tempst P, et al.

RNA polymerase II elongator holoenzyme is composed of two discrete subcomplexes. J

Biol Chem 2001;276:32743-49.

[38] Collum RG, Brutsaert S, Lee G, Schindler C. A Stat3-interacting protein (StIP1) regulates

cytokine signal transduction. Proc Natl Acad Sci U S A 2000;97:10120-25.

[39] Wang T, Niu G, Kortylewski M, Burdelya L, Shain K, Zhang S, et al. Regulation of the innate

and adaptive immune responses by Stat-3 signaling in tumor cells. Nat Med 2004;10:48-54.

[40] Wang Y, Shen Y, Wang S, Shen Q, Zhou X. The role of STAT3 in leading the crosstalk

between human cancers and the immune system. Cancer Lett 2018;415:117-28.

[41] Lee JS, Choi HJ, Baek SH. Sumoylation and Its Contribution to Cancer. Adv Exp Med Biol

2017;963:283-98.

[42] Rabellino A, Carter B, Konstantinidou G, Wu SY, Rimessi A, Byers LA, et al. The SUMO E3-

ligase PIAS1 regulates the tumor suppressor PML and its oncogenic counterpart PML-

RARA. Cancer Res 2012;72:2275-84.

[43] Ni L, Lu J. Interferon gamma in cancer immunotherapy. Cancer Med 2018; 7:4509-16.

[44] Qing Y, Stark GR. Alternative activation of STAT1 and STAT3 in response to interferon-

gamma. J Biol Chem 2004;279:41679-85.

[45] Prium R, Welch R. LocusZoom: regional visualization of genome-wide association scan

results. Bioinformatics 2010;26:2336-37.

[46] Chanda A, Chan A, Deng L, Kornaga EN, Enwere EK, Morris DG et al. Identification of the

SUMO E3 ligase PIAS1 as a potential survival biomarker in breast cancer. PLoS ONE

2017;12:e0177639.

[47] Berland L, Heeke S, Humbert O, Macoco A, Long-Mira E, Lasalle S, et al. Current views on

tumor mutation burden in patients with non-small cell lung cancer treated by immune

checkpoint inhibitors. J Thoracic Dis 2019;11:71-80.

[48] Graves E, Maity A, Le Q. The Tumor Microenvironment in Non-small Cell Lung Cancer.

Semin Radiat Oncol 2010;20:156-163.

Table 1. Associations of three significant SNPs with overall survival of patients with NSCLC in both discovery and validation datasets from two Downloaded from previously published GWASs Author manuscriptshavebeenpeerreviewedandacceptedforpublicationbutnotyetedited. Author ManuscriptPublishedOnlineFirstonJune3,2020;DOI:10.1158/1055-9965.EPI-19-1450 SNPs* Allelea Gene PLCO (n=1185) HLCS (n=984) Combined-analysis b b c c d 2 e e FDR BFDP EAF HR (95% CI) P EAF HR (95% CI) P P het I HR (95% CI) P cebp.aacrjournals.org rs7242481 G>A ELP2 0.50 0.56 0.35 0.86 (0.77-0.95) 0.004 0.36 0.83 (0.74-0.93) 0.002 0.658 0 0.85 (0.78-0.92) 3.02x10-5 rs11071978 T>A PIAS1 0.50 0.72 0.36 0.86 (0.78-0.96) 0.006 0.34 0.82 (0.72-0.92) 0.001 0.529 0 0.84 (0.77-0.91) 2.34x10-5 rs1049493 T>C PIAS1 0.50 0.69 0.45 0.87 (0.79-0.96) 0.006 0.45 0.87 (0.78-0.96) 0.009 0.989 0 0.87 (0.81-0.93) 1.24x10-4

*The other 45 SNPs in high LD (r2 >0.8) with these three SNPs are presented in Supplementary Table 5 and Supplementary Table 6. Abbreviations: SNPs, single-nucleotide polymorphisms; NSCLC, non-small cell lung cancer; GWAS, genome-wide association study; PLCO, the Prostate, on October 1,2021. ©2020 American Association forCancer Research. Lung, Colorectal and Ovarian cancer screening trial; HLCS, Harvard Lung Cancer Susceptibility Study; EAF, effect allele frequency; HR, hazards ratio; CI, confidence interval; FDR, false discovery rate; BFDP, Bayesian false discovery probability; LD, linkage disequilibrium; aEffect/reference allele; bAdjusted for age, sex, stage, histology, smoking status, chemotherapy, radiotherapy, surgery, PC1, PC2, PC3 and PC4; cAdjusted for age, sex, stage, histology, smoking status, chemotherapy, radiotherapy, surgery, PC1, PC2 and PC3; d Phet: P value for heterogeneity by Cochrane’s Q test; eMeta-analysis in the fix-effects model.

Table 2. Two indenpendent SNPs in a multivariate Cox proportional hazards regression analysis with adjustment for other Downloaded from covariates and 15 previously published SNPs for NSCLC in the PLCO Trial Author manuscriptshavebeenpeerreviewedandacceptedforpublicationbutnotyetedited. a a b b

Variables Category Frequency HR (95% CI) P HR (95% CI) P Author ManuscriptPublishedOnlineFirstonJune3,2020;DOI:10.1158/1055-9965.EPI-19-1450 Age Continuous 1185 1.03 (1.02-1.05) <0.0001 1.04 (1.02-1.05) <0.0001 Sex Male 698 1.00 1.00 cebp.aacrjournals.org Female 487 0.80 (0.69-0.93) 0.004 0.78 (0.67-0.91) 0.002 Smoking status Never 115 1.00 1.00 Current 423 1.67 (1.26-2.26) 0.0004 1.94 (1.44-2.62) <0.0001 Former 647 1.64 (1.25-2.16) 0.0004 1.90 (1.42-2.52) <0.0001 Histology Adenocarcinoma 577 1.00 1.00 on October 1,2021. ©2020 American Association forCancer Research. Squamous cell 285 1.20 (0.99-1.45) 0.057 1.25 (1.03-1.51) 0.025 others 323 1.32 (1.11-1.56) 0.002 1.37 (1.15-1.63) 0.0006 Tumor stage I-IIIA 655 1.00 1.00 IIIB-IV 528 2.94 (2.42-3.58) <0.0001 3.11 (2.55-3.79) <0.0001 Chemotherapy No 639 1.00 1.00 Yes 538 0.58 (0.48-0.69) <0.0001 0.58 (0.48-0.70) <0.0001 Radiotherapy No 762 1.00 1.00 Yes 415 0.95 (0.81-1.12) 0.526 0.94 (0.80-1.12) 0.497 Surgery No 637 1.00 1.00 Yes 540 0.21 (0.17-0.28) <0.0001 0.20 (0.15-0.26) <0.0001 ELP2 rs7242481 G>A GG/GA/AA 495/544/146 0.86 (0.78- 0.96) 0.007 0.86 (0.77- 0.96) 0.007 PIAS1 rs1049493 T>C TT/TC/CC 368/579/248 0.87 (0.79- 0.97) 0.009 0.89 (0.80- 0.99) 0.024

Abbreviations: SNP: single-nucleotide polymorphisms; NSCLC, non-small cell lung cancer; PLCO, the Prostate, Lung, Colorectal and Ovarian cancer screening trial; HLCS, Harvard Lung Cancer Susceptibility Study; HR: hazards ratio; CI: confidence interval. aAdjusted for age, sex, tumor stage, histology, smoking status, chemotherapy, radiotherapy, surgery, and PC1, PC2, PC3, PC4; bOther 15 published SNPs were included for further adjustment: five SNPs in PMID: 27557513; one SNP in PMID: 29978465; two SNPs in PMID: 30259978; two SNPs in PMID: 26757251; three SNPs in PMID: 30650190; and two SNPs in PMID: 30989732.

Table 3. Associations between the number of protective alleles of two independent SNPs with OS and DSS of NSCLC in the Downloaded from PLCO Trial

b b Author manuscriptshavebeenpeerreviewedandacceptedforpublicationbutnotyetedited. Genotypes/Alleles Frequencya OS DSS Death (%) HR (95% CI) P Death (%) HR (95% CI) P Author ManuscriptPublishedOnlineFirstonJune3,2020;DOI:10.1158/1055-9965.EPI-19-1450 ELP2 rs7242481 G>A GG 491 336 (68.43) 1.00 306 (62.32) 1.00

cebp.aacrjournals.org GA 538 361 (67.10) 0.85 (0.73-0.99) 0.033 318 (59.11) 0.84 (0.72-0.99) 0.033 AA 146 92 (63.01) 0.74 (0.59-0.94) 0.012 85 (58.22) 0.76 (0.59-0.97) 0.026 Trend test 0.004 0.009 Dominant GG 491 336 (68.43) 1.00 306 (62.32) 1.00 GA+AA 684 453 (66.23) 0.82 (0.71-0.95) 0.008 403 (58.92) 0.82 (0.81-0.96) 0.011 PIAS1 rs1049493 T>C on October 1,2021. ©2020 American Association forCancer Research. TT 356 250 (70.22) 1.00 225 (63.20) 1.00 TC 571 377 (66.02) 0.92 (0.78-1.08) 0.303 334 (58.49) 0.91 (0.76-1.08) 0.285 CC 248 162 (65.32) 0.75 (0.61-0.91) 0.005 150 (60.48) 0.78 (0.63-0.97) 0.022 Trend test 0.006 0.023 Dominant TT 356 250 (70.22) 1.00 225 (63.20) 1.00 TC+CC 819 539 (65.81) 0.86 (0.74-1.00) 0.053 484 (59.10) 0.87 (0.74-1.02) 0.082 NPGc 0 398 274 (68.84) 1.00 249 (62.06) 1.00 1 622 415 (66.72) 0.94 (0.81-1.10) 0.424 371 (59.65) 0.95 (0.81- 1.12) 0.566 2 155 100 (64.52) 0.62 (0.49-0.79) <0.0001 91 (58.71) 0.64 (0.50- 0.83) 0.001 Trend test 0.0004 0.002 Dichotomized NPG 0-1 1020 689 (67.6) 1.00 618 (60.59) 1.00 2 155 100 (64.5) 0.64 (0.52-0.80) <0.0001 91 (58.71) 0.66 (0.53-0.83) 0.0004 NPAd 0 140 100 (71.43) 1.00 88 (62.86) 1.00 1 431 295 (68.45) 1.03 (0.82-1.30) 0.796 269 (62.41) 1.08 (0.84-1.37) 0.561 2 379 248 (65.44) 0.85 (0.68-1.08) 0.189 222 (58.58) 0.89 (0.69-1.14) 0.355 3 192 120 (62.50) 0.74 (0.56-0.97) 0.028 110 (57.29) 0.79 (0.59-1.05) 0.103 4 33 21 (63.64) 0.51 (0.31-0.82) 0.005 19 (57.58) 0.53 (0.32-0.81) 0.013 Trend test <0.0001 0.0006 Dichotomized NPA 0-1 571 395 (69.18) 1.00 357 (62.52) 1.00 2-4 604 389 (64.40) 0.77 (0.67-0.89) 0.0004 351 (58.11) 0.79 (0.68- 0.91) 0.002 Abbreviations: SNP, single nucleotide polymorphism; NSCLC, non-small cell lung cancer; PLCO, Prostate, Lung, Colorectal and Ovarian cancer screening trial; HR, hazards ratio; CI, confidence interval; OS, overall survival; DSS, disease-specific survival. NPG, number of

Downloaded from protective genotypes; NPA: number of protective alleles. a10 missing date were excluded; Author manuscriptshavebeenpeerreviewedandacceptedforpublicationbutnotyetedited. bAdjusted for age, sex, smoking status, histology, tumor stage, chemotherapy, surgery, radiotherapy and principal components; Author ManuscriptPublishedOnlineFirstonJune3,2020;DOI:10.1158/1055-9965.EPI-19-1450 cProtective genotypes were ELP2 rs7242481 GA+AA and PIAS1 rs1049493 TC+CC; dProtective alleles were ELP2 rs7242481_A and PIAS1 rs1049493_C. cebp.aacrjournals.org

on October 1,2021. ©2020 American Association forCancer Research. Author Manuscript Published OnlineFirst on June 3, 2020; DOI: 10.1158/1055-9965.EPI-19-1450 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Figure Legends

Figure 1: Study flowchart: The overall procedures of the present study.

Abbreviations: SNPs, single-nucleotide polymorphisms; MAF: minor allelic frequency; HWE:

Hardy-Weinberg Equilibrium; PLCO, The Prostate, Lung, Colorectal and Ovarian Cancer

Screening Trial; HLCS, the Harvard Lung Cancer Susceptibility Study; NSCLC, non-small cell lung cancer; ELP2, elongator acetyltransferase complex subunit 2; PIAS1, protein inhibitor of activated STAT1.

Figure 2: Prediction of survival with combined protective alleles

Abbreviations: OS, overall survival; DSS, disease-specific survival; NPA: number of protective alleles; PLCO, The Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial.

Kaplan-Meier (KM) survival curves for NSCLC patients of two validated SNPs and combined protective alleles in the PLCO trial. (a) by 0, 1 and 2 protective alleles in OS, (b) by 0, 1 and 2 protective alleles in DSS, (c) Dichotomized groups of the NPA into 0-1 and 2-4 in OS, and (d)

Dichotomized groups of the NPA into 0-1 and 2-4 in DSS from the PLCO trial.

Figure 3: Correlation of genotypes with the mRNA expression levels of the corresponding genes

Abbreviations: eQTL, expression quantitative trait loci; GTEx, Genotype-Tissue Expression project; ELP2, elongator acetyltransferase complex subunit 2; PIAS1, protein inhibitor of activated STAT1. eQTL analysis of the independent and significant SNPs associated with NSCLC survival. rs7242481 A allele is associated with a higher mRNA expression of ELP2 (a) in 373 European samples from the 1000 Genomes Project and (c) in 369 genotyped whole blood samples, but not in (b) 383 genotyped normal lung tissues from the GTEx project; rs1049493 C allele was not significantly associated with an altered mRNA expression of PIAS1 (d) in 373 European

samples from the 1000 Genomes Project, but is significantly associated with a decreased mRNA expression of PIAS1 (e) in 383 normal lung tissue and (f) 369 whole blood samples from the GTEx project.

Downloaded from cebp.aacrjournals.org on October 1, 2021. © 2020 American Association for Cancer Research. Downloaded from Author manuscriptshavebeenpeerreviewedandacceptedforpublicationbutnotyetedited. Author ManuscriptPublishedOnlineFirstonJune3,2020;DOI:10.1158/1055-9965.EPI-19-1450 Figure 1Figure 1 cebp.aacrjournals.org on October 1,2021. ©2020 American Association forCancer Research. Downloaded from Figure 2 Author manuscriptshavebeenpeerreviewedandacceptedforpublicationbutnotyetedited. Author ManuscriptPublishedOnlineFirstonJune3,2020;DOI:10.1158/1055-9965.EPI-19-1450 cebp.aacrjournals.org on October 1,2021. ©2020 American Association forCancer Research. Downloaded from Figure 3 Author manuscriptshavebeenpeerreviewedandacceptedforpublicationbutnotyetedited. Author ManuscriptPublishedOnlineFirstonJune3,2020;DOI:10.1158/1055-9965.EPI-19-1450 cebp.aacrjournals.org on October 1,2021. ©2020 American Association forCancer Research. Author Manuscript Published OnlineFirst on June 3, 2020; DOI: 10.1158/1055-9965.EPI-19-1450 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Novel variants of ELP2 and PIAS1 in the interferon gamma signaling pathway are associated with non-small cell lung cancer survival

Yu Chen Zhao, Dongfang Tang, Sen Yang, et al.

Cancer Epidemiol Biomarkers Prev Published OnlineFirst June 3, 2020.

Updated version Access the most recent version of this article at: doi:10.1158/1055-9965.EPI-19-1450

Supplementary Access the most recent supplemental material at: Material http://cebp.aacrjournals.org/content/suppl/2020/06/03/1055-9965.EPI-19-1450.DC1

Author Author manuscripts have been peer reviewed and accepted for publication but have not yet been Manuscript edited.

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cebp.aacrjournals.org/content/early/2020/06/03/1055-9965.EPI-19-1450. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.